SALGINTR: development and early evaluation of a low-cost cloud-based digital participatory surveillance system for influenza-like Illness among physicians in Türkiye — a proof of concept

Date

2026-12-01

Author

ÖNTAŞ, ERAY
Güçlü, Hasan
Aydın Son, Yeşim

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

982
views

0
downloads

Cite This

Background: Participatory surveillance complements sentinel influenza monitoring in high-income settings, yet no such platform has targeted physicians as both participants and clinical sentinels. We established SALGINTR, a digital, physician-driven participatory surveillance system for influenza-like illness (ILI), assembling an automated data collection and analysis pipeline from freely available cloud-based tools at a total infrastructure cost of approximately US $100. Methods: Physicians across Türkiye were recruited via snowball sampling to provide weekly self-reports of ILI symptoms through an online platform (www.salgin.com.tr) during epidemiological weeks 40–52 of the 2025–26 influenza season. ECDC ILI case definition was applied, and independent ILI episodes were ascertained using a clinical-episode rule whereby consecutive symptomatic weeks attributed to the same agent were counted as a single episode. Temporal concordance with national sentinel surveillance was assessed by Pearson and Spearman correlation with moving-block bootstrap confidence intervals, lead–lag cross-correlation, and Bland–Altman agreement. Epidemic detection was evaluated using a panel of six aberration detection algorithms (EARS C1–C3, negative-binomial CUSUM, EWMA, empirical 95th percentile). Results: Of 243 registered physicians, 197 (analytical cohort) contributed 1,484 respondent-weeks of observation. Cumulative ILI incidence was 50.8% (100/197), with 156 independent episodes identified from 190 symptomatic person-weeks using a clinical-episode definition. Season-level viral positivity was similar in the two streams (SALGINTR 17.3%, 95% CI 11.7–24.2, 27/156; sentinel 14.5%, 95% CI 12.8–16.2, 247/1,709). Weekly SALGINTR viral-systemic positivity was temporally concordant with sentinel surveillance under the pre-specified causal three-week moving average (Pearson r = 0.575, 95% CI 0.35 to 0.90; Spearman ρ = 0.747; pre-epidemic W40–W49 r = 0.886, 95% CI 0.38 to 0.94). Bland–Altman analysis indicated no significant systematic bias (mean difference + 4.6 pp; 95% limits of agreement − 18.3 to + 27.4 pp). Neither system produced any alarm during the ten pre-epidemic weeks, and alarm agreement across the window was substantial (Cohen’s κ = 0.755; specification-curve median κ = 0.639 across 72 analytical choices). Lead–lag analysis identified no timing advantage in either direction (peak lag − 1 week; bootstrap CI − 2 to + 2 weeks). Conclusions: SALGINTR demonstrates that a physician-driven digital participatory surveillance system can generate ILI signals concordant with established sentinel surveillance, at minimal cost. This proof-of-concept supports integration of participatory surveillance as a complementary component within mosaic respiratory disease surveillance frameworks in middle-income settings.

URI

https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=105047600070&origin=inward
https://hdl.handle.net/11511/120265

Collections

Graduate School of Informatics, Article